Molecular mechanism of c‐Myc and PRPS1/2 against thiopurine resistance in Burkitt's lymphoma

Abstract Patients with relapsed/refractory Burkitt's lymphoma (BL) have a dismal prognosis. Current research efforts aim to increase cure rates by identifying high‐risk patients in need of more intensive or novel therapy. The 8q24 chromosomal translocation of the c‐Myc gene, a main molecular marker of BL, is related to the metabolism by regulating phosphoribosyl pyrophosphate synthetase 2 (PRPS2). In our study, BL showed significant resistance to thiopurines. PRPS2 homologous isoenzyme, PRPS1, was demonstrated to play the main role in thiopurine resistance. c‐Myc did not have direct effects on thiopurine resistance in BL for only driving PRPS2. PRPS1 wild type (WT) showed different resistance to 6‐mercaptopurine (6‐mp) in different metabolic cells because it could be inhibited by adenosine diphosphate or guanosine diphosphate negative feedback. PRPS1 A190T mutant could dramatically increase thiopurine resistance in BL. The interim analysis of the Treatment Regimen for Children or Adolescent with mature B cell non‐Hodgkin's lymphoma in China (CCCG‐B‐NHL‐2015 study) confirms the value of high‐dose methotrexate (MTX) and cytarabine (ARA‐C) in high‐risk paediatric patients with BL. However, there remains a subgroup of patients with lactate dehydrogenase higher than four times of the normal value (4N) for whom novel treatments are needed. Notably, we found that the combination of thiopurines and the phosphoribosylglycinamide formyltransferase (GART) inhibitor lometrexol could serve as a therapeutic strategy to overcome thiopurine resistance in BL.


| INTRODUC TI ON
Burkitt's lymphoma (BL) is one of the highly invasive mature B cell malignant tumours, and its incidence rate is about 4.5/1 000 000.
The genetic characteristic of BL is the 8q24 chromosomal translocation of the c-Myc gene which is the main marker gene of BL. 1 c-Myc gene plays a crucial role in many important physiological activities such as cell division and proliferation, metabolism, cell cycle and apoptosis. 2,3 It has been found that the activated c-Myc gene can induce the tumour metabolism. Therefore, the therapy of targeting c-Myc is a trend in the treatment of high-risk and relapsed or progressive BL. 4 Davide Ruggero demonstrated that c-Myc could regulate the metabolic changes of tumour cells by regulating the rate-limiting enzyme phosphoribosyl pyrophosphate synthetase 2 (PRPS2). 5 PRPS1 and PRPS2 which catalyse the same biochemical reaction 6 are critical rate-limiting purine biosynthesis enzymes. However, PRPS1 has higher enzymatic and stronger allosteric activity than PRPS2. 7 PRPS1 can also be significantly inhibited by adenosine diphosphate (ADP) and guanosine diphosphate (GDP) negative feedback. 8 It was reported that negative feedback-defective PRPS1 mutants could drive thiopurine resistance in relapsed childhood acute lymphoblastic leukaemia (ALL). 9 It is unknown whether c-Myc and PRPS2 are associated with the thiopurine resistance.
In the current study, we compared the characteristic of chemotherapy drug resistance in BL with that in leukaemia and dis-

| Cell construction
Phosphoribosyl pyrophosphate synthetase 1 WT, PRPS2 WT, c-Myc coding DNA sequences were cloned into a pGV287 green fluorescent protein (GFP) vector. The PRPS1 A190T coding region was cloned into pGV303 GFP vector (Shanghai GeneChem). To produce the lentivirus, each expression vector was transfected into 293T cells with second-generation lentiviral packaging plasmids pMD2. G and psPX2 using the PolyExpress transfection reagent (Excellgen).
48 and 72 hours after transfection, we harvested the culture medium, incubated with Lenti-X concentrator (Clontech Laboratories), and centrifuged to obtain the concentrated lentivirus. The 293T cells were infected with the lentiviruses in the presence of 6 μg/mL polybrene (Sigma-Aldrich) for 24 hours. Overexpression was confirmed by Western blot.

| Cell viability and apoptosis
Cell viability was determined by using the CellTiter-Glo Luminescent kit (Promega, Madison, Wisconsin, USA) according to the manufacturer's instructions. Cells were seeded in 96-well plates (10 000 cells per well) and treated for 72 hours with serially diluted drugs.
CellTiter-Glo reagents (50 μL) was added to each well and mixed for 10 minutes before the luminescence was measured on a microplate reader (BioTek, Vermont, USA). Apoptosis was measured by staining with annexin V-APC and Propidium Iodide (PI)-phycoerythrin (PE) (Annexin V-APC Apoptosis Detection kit, BD Pharmingen), followed by flow cytometry on a FACS flow cytometer (BD, Canto II). All experiments were performed in triplicate, and results were calculated as the mean ± SD.

| Metabolite flux
Cells were cultured for 48 hours in RPMI 1640, then harvested and pelleted. The reaction was quenched in cold 80% methanol, cells were centrifuged at 1200 g for 10 minutes, and metabolites in the supernatant were analysed by LC-MS. The relative concentrations were defined according to the standard curve of compounds dissolved in 80% methanol without correcting for cell matrix effect.

| Reverse transcription-PCR analysis
MSH2 cDNA fragments containing the region between exon 9 and exon 16 were amplified by reverse transcription-PCR with TaKaRa RNA PCR Kit (AMV) Ver.3.0 using Nalm6 and Reh total RNA as template. The primer sets were as follows: Forward: 5'-AAGCTGATTGGGTGTGGTCG-3'.

| Statistical analysis
All experiments were repeated three times. Data are presented as mean ± SD. Two-tailed Student's t tests were performed, and P < 0.05 was considered statistically significant.

| Drug resistance to thiopurines related to high c-Myc expression as well as nucleotide metabolic level
In BL, some patients do not respond to initial therapy or relapse after the standard therapy, which leads to poor prognosis. The mechanisms underlying BL chemoresistance remain poorly defined.
To uncover the intrinsic mechanism, we used several cell lines of

| Direct manipulation of c-Myc has no effects on thiopurine drug resistance
To confirm that c-Myc was related to thiopurine drug resistance, we Then, we knocked down c-Myc in high metabolic BL cell lines ( Figure 2E, Figure S2A) and low metabolic Vocb6 ALL cell line ( Figure 2K), whereas we overexpressed c-Myc in low metabolic ALL cell lines ( Figure 2H, Figure S2C). The c-Myc knockdown had little effects on thiopurine drug resistance ( Figure 2F,l, Figure S2B), and the c-Myc overexpressed cell lines conferred thiopurine resistance similarly to control cells as well ( Figure 2I, Figure S2D). Then, we detected the level of nucleotide metabolism in c-Myc regulated BL and ALL cell lines mentioned above. The regulation of c-Myc also had faint influences on purine metabolism ( Figure 2G,j,m, Figure S2E). Therefore, our results suggested that direct regulation of c-Myc did not affect thiopurine drug resistance. What's more, PRPS2 may have no function on thiopurine resistance, and some other molecules regulating the purine metabolism remain uncovered.

| PRPS1 plays the key role in the thiopurine resistance regulated by the level of purine metabolites such as IMP, ADP and GDP
On the basis of the role of PRPS1 and PRPS2 in purine biosynthesis, we examined whether the regulation of PRPS1 or PRPS2 had the same thiopurine drug resistance in different metabolic cell lines. We established ALL and BL cell lines infected with PRPS1 wild type (1-wt), PRPS2 wild type (2-wt), PRPS1 knocked out (1-ko) and PRPS2 knocked out (2-ko) retroviruses ( Figure 3A,D, Figure S3A,C,E). In low purine metabolic ALL cell lines, the overexpression of PRPS1 obviously increased the level of thiopurine drug resistance, while overexpressed PRPS2 did not obviously change the 6-mpIC50 ( Figure 3B, Figure S3B).
Moreover, knocking outPRPS1 in BL and ALL cell lines could dramatically decrease the 6-mp drug resistance ( Figure 3B,E, Figure S3B,D,F).
We observed that Vocb6PRPS1-WT cell line had higher purine metabolism than Vocb6PRPS2-WT cell line as well ( Figure 3C). As shown in Figure 3E, Figure S3D,F, overexpression of PRPS1 or PRPS2 in high metabolic cell lines showed little effects on thiopurine drug resistance.
Moreover, similar to the ALL cell lines, knocking out PRPS1 in BL cell line could significantly decrease the IC50 of 6-mp. Knockout of PRPS2 only slightly decreased the IC50 of 6-mp ( Figure 3E, Figure S3D,F). In addition, we investigated that PRPS2 WT showed no difference in purine metabolism with PRPS1 WT Namalwa cells, consistent with 6-mp resistance data mentioned above ( Figure 3F). The results suggested that PRPS1 played the main role in the thiopurine resistance, while PRPS2 did not affect the drug sensitivity significantly.
Nalm6 is a MSH2-deficient precursor B cell lymphoblastic leukaemia (BCP-ALL) cell line 12 ( Figure S4A,B). MSH2-deficient cell lines are highly sensitive to thiopurine drugs. 13 Even though we overexpressed and knocked out PRPS1 and PRPS2 in the Nalm6 cell line ( Figure S4C), the regulation of PRPS1 or PRPS2 slightly affected the thiopurine drug resistance ( Figure S4D). Therefore, our data suggested that mutations which affected the stability of DNA could overcome the influence of regulating PRPS.
To understand why PRPS1 had different effects on thiopurine resistance in different metabolic BL and ALL cell lines, we examined the influence of nucleic acid metabolite ADP and HX on thiopurine resistance. First, we detected the sensitivity of HX and ADP in both BL and ALL cell lines to select appropriate concentrations ( Figure   S1B) and then tested the thiopurine resistance after adding HX or ADP. In all cancer cells, adding HX highly increased the resistance to thiopurine drugs in the same level ( Figure 3G). Therefore, in BL, HX was significantly higher than that in ALL cell lines, which was one reason for drug resistance to 6-mp ( Figure 1F). Adding ADP significantly led to higher sensitivity of thiopurine drugs in all PRPS1 overexpressed cells, while PRPS2 overexpressing was resistant to the effects of ADP ( Figure 3H). In BL, ADP and GDP were also significantly higher than those in ALL ( Figure 1F). Therefore, overexpression of PRPS1 could not exhibit the enzymatic activity or affect the IC50 of 6-mp. Our results suggested that high HX and ADP/GDP manipulated PRPS1 activity and the 6-mp IC50. As a result, it implicates that PRPS1, not PRPS2, is the key factor in thiopurine resistance.

| PRPS1 A190T is a resistant mutation of thiopurine by avoiding ADP or GDP inhibition
Activating PRPS1 A190T mutant can escape nucleotide feedback inhibition, so PRPS1 A190T mutant can be a resistant mutation of 6-mp in relapsed childhood ALL. 9 To know whether the PRPS1 A190T mutant unlike PRPS1 WT showed the same drug resistance in BL as in ALL, we established the PRPS1 A190T mutation in BL and ALL cell lines ( Figure 4A). We found that the PRPS1 A190T mutant presented  Figure 4B). In agreement with this, the purine metabolic level of A190T mutant cells was much higher ( Figure 4C). Moreover, the relative thiopurine resistance in high metabolic cell lines was much lower than that in low metabolic cell lines ( Figure 4B). Then, we tested the different concentrations on the 6-mp IC50 of PRPS1-A190T mutant cells, and we found that the 6-mp IC50 of PRPS1-A190T mutant cells decreased at increasing concentrations of ADP ( Figure 4D). PRPS1-A190T in low metabolic cell lines could escape higher concentration of ADP than PRPS1-A190T in high metabolic cell lines ( Figure 4D). These data suggest that the concentration of ADP/GDP is important to the thiopurine resistance. When the concentration is high enough, ADP/ GDP can also inhibit the function of PRPS1-A190T. In a word, the thiopurine resistance of A190T mutations is limited.

| The combination of thiopurines and lometrexol is benefit to BL treatment
It has been reported that HD MTX and Ara-C are the major drugs in BL. 14 Figure 5B). All these clinical data suggest that the therapy research targeting for its metabolic pathway is clinically reasonable and applicable.
Lometrexol (also called DDATHF) which is one of the GART inhibitors has been tested as an antitumour drug. 15 We firstly detected the IC50 of lometrexol in BL and ALL cells to choose the appropriate concentration (Fig s1B) and then did combination studies. According to our results, the lometrexol treatment dramatically decreased the thiopurine resistance in BL cells to ALL cells level ( Figure 5C). Next, we tested the nucleotide metabolic level after adding the combination of 6-mp and lometrexol in BL and ALL cells. Compared to adding thiopurine only, the combination of thiopurine and lometrexol showed obviously lower F I G U R E 4 PRPS1 A190T is a resistant mutation of thiopurine by avoiding ADP or GDP inhibition. A, Western blot of the PRPS1 A190T mutant in Burkitt's lymphoma (Namalwa, Raji) and ALL cell lines (Molt4, Vocb6). β-Actin was used as a loading control. B, The 6-mpIC50 of PRPS1 A190T mutant Namalwa, Raji, Molt4 and Vocb6. Unregulated cells as control. *:P < 0.05; two-tailed Student's t tests. C, Heatmap showing nucleotide metabolism of PRPS1 A190T mutant in Namalwa and Vocb6. D, The 6-mp IC50 of PRPS1 A190T mutant Vocb6 and Namalwa at increasing concentrations of ADP (µmol/L).*: P < 0.05; two-tailed Student's t tests. ADP, adenosine diphosphate; ALL, acute lymphoblastic leukaemia; GDP, guanosine diphosphate; PRPS1, phosphoribosyl pyrophosphate synthetase 1

TA B L E 2
The stage and treatment protocol in the CCCG-B-NHL-2015 metabolic levels in BL and ALL cells ( Figure 5D). What's more, we observed that the lometrexol treatment could also decrease the thiopurine resistance in PRPS1 A190T mutant cells ( Figure S1C).
The IC50 of 6-mp in normal human cells like HF cells was much higher than the concentration we used in cancer cells ( Figure S1D).
Simultaneously, the IC50 of the combination treatment in HF cells was much higher than that in cancer cells ( Figure S1D). Therefore, this combination did not have any added toxic effects on normal cells. These data suggested that the combination of 6-mp and lometrexol might serve as a therapeutic strategy to overcome thiopurine drug resistance in BL. to thiopurine resistance was much lower than that in ALL cells.

| D ISCUSS I ON
These data suggest that the PRPS1 A190T mutant can be inhibited if the level of nucleotide metabolism is high enough. F I G U R E 5 The combination of thiopurines and lometrexol is benefit to Burkitt's lymphoma treatment. A, EFS at 2 y for mature B cell non-Hodgkin lymphoma paediatric patients in the two consecutive studies CCCG-BNHL-2015 (n = 314, interim analysis, unpublished data) and CCCG-BNHL-2010 (n = 104). P = 0.003. b, The 2-year EFS of patients with different levels of LDH.N: the normal value. *P < 0.05; two-tailed Student's t tests. C, The cell sensitivity to 6-mp after treating with the 5 ng/mL GART inhibitor lometrexol. *P < 0.05; two-tailed Student's t tests. D, Heatmap showing nucleotide metabolism after treating with 10 µg/mL 6-mp or the combination of 10 µg/mL 6-mp and 10 ng/mL lometrexol in Namalwa and Vocb6. EFS, event-free survival; LDH, lactate dehydrogenase

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest with the contents of this article.

AUTH O R CO NTR I B UTI O N S
This study was conceived by YJ G and YX L.; YJ G and YX L. designed the study; T. L., LL S., YL, ZY Z, YW Z, YY, SQ W and SS L performed the experiments; T. L. analysed and interpreted the data; YJ G reviewed the manuscript; and TL and YX L. wrote the paper with comments from all authors. All authors read and approved the final manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data supporting the findings of this study are available from the corresponding author on request.